ABSTRACTSamples of speech suitable for sociolinguistic analysis may be sought in several ways. Interviews (either formal or informal), and tape-recorded group sessions, are the methods most used currently. In research on a specific variable, the historical present tense (HP), none of these methods proved neutral or adequate. Although the historical present tense is very widely used in conversational narratives, its occurrence within interviews is so infrequent as to be striking. An explanation was found in the way in which the interview has a specific known place as a speech event in the culture of those whose speech was being studied. The so-called spontaneous interview does not have such a place, and for that very reason is even less satisfactory a source of data. The notion of natural speech is taken as properly equivalent to that of appropriate speech; as not equivalent to unselfconscious speech; and as observable easily, and often best, by simple techniques of participation. (Sociolinguistic methodology; speech events, interviews, observation, natural speech; United States English).